METHOD AND SYSTEM FOR VIDEO DECODING BY MEANS OF A GRAPHIC PIPELINE, COMPUTER PROGRAM PRODUCT THEREFOR

• US 20090160866A1
• Filed: 12/22/2008
• Published: 06/25/2009
• Est. Priority Date: 06/22/2006
• Status: Active Grant

• Current Assignee: STMicroelectronics SRL (STMicroelectronics NV)
• Original Assignee: STMicroelectronics SRL (STMicroelectronics NV)
• Inventors: Lavigna, Daniele, Pau, Danilo, Borneo, Antonio Maria
• Granted Patent: US 8,269,781 B2
• Time in Patent Office: Days
• US Class Current: 345/503
• CPC Class Codes: G06T 9/001 Model-based coding, e.g. wire frame
• CPC Class Codes: H04N 19/27 involving both synthetic and natural picture components, e.g. synthetic natural hybrid coding [SNHC]
• CPC Class Codes: H04N 19/42 characterised by implementation details or hardware specially adapted for video compression or decompression, e.g. dedicated software implementation H04N19/635 takes precedence
• CPC Class Codes: H04N 19/436 using parallelised computational arrangements
• CPC Class Codes: H04N 19/44 Decoders specially adapted therefor, e.g. video decoders which are asymmetric with respect to the encoder

Abstract

A system for decoding a stream of compressed digital video images comprises a graphics accelerator for reading the stream of compressed digital video images, creating, starting from said stream of compressed digital video images, three-dimensional scenes to be rendered, and converting the three-dimensional scenes to be rendered into decoded video images. The graphics accelerator is preferentially configured as pipeline selectively switchable between operation in a graphics context and operation for decoding the stream of video images. The graphics accelerator is controllable during operation for decoding the stream of compressed digital video images via a set of application programming interfaces comprising, in addition to new APIs, also standard APIs for operation of the graphics accelerator in a graphics context.

Claims

1. A method for decoding a stream of compressed digital video images, comprising:

reading said stream of compressed digital video images using a graphics accelerator;

creating, via said graphics accelerator, starting from said stream of compressed digital video images, three-dimensional scenes to be rendered; and

converting said three-dimensional scenes to be rendered into decoded video images.

2. The method according to claim 1, further comprising configuring said graphics accelerator as a pipeline.

3. The method according to claim 1, further comprising providing an enumerator for switching said graphics accelerator between operation in a graphics context and operation for decoding said stream of video images.

4. The method according to claim 1, further comprising controlling said graphics accelerator during operation for decoding said stream of compressed digital video images via a set of application programming interfaces.

5. The method according to claim 4 wherein said set of application programming interfaces comprises standard APIs for operation of said graphics accelerator in a graphics context.

6. The method according to claim 1, further comprising submitting said stream of compressed digital video images to at least one operation chosen in a group consisting of:

header decoding;

inverse variable-length coding;

zigzag scanning;

obtaining of quantized data regarding blocks of the stream; and

decoding of motion vectors of the stream.

7. The method according to claim 1, further comprising deriving, from said stream of compressed digital video images, data on frame dimensions and frame number, type of images, and type of macroblocks.

8. The method according to claim 1 wherein creating said three-dimensional scenes to be rendered includes creating meshes of primitives.

9. The method according to claim 8, wherein creating said meshes of primitives includes creating meshes of two-dimensional geometrical figures.

10. The method according to claim 8 wherein creating said meshes of primitives includes containing the meshes of primitives in a vertex buffer object.

11. The method according to claim 8, further comprising performing a rasterization operation that returns, starting from said meshes of primitives, a set of pixels representing prediction error frames of said decoded video image.

14. The method according to claim 1 wherein creating said three-dimensional scenes to be rendered includes creating said three-dimensional scenes to be rendered via inverse quantization and inverse discrete cosine transform.

15. The method according to claim 1 wherein converting said three-dimensional scenes to be rendered into decoded video images comprises combining said three-dimensional scenes to be rendered with motion data of said compressed digital video images and stored reference images, and applying motion compensation.

16. The method according to claim 15, further comprising supplying said converting operation with signs of the motion data separately from the motion data.

17. The method according to claim 16 wherein:

said converting is performed via a texture unit provided with a channel for transmission of transparency data; and

supplying said converting operation with the signs includes using said channel for transmission of transparency data to transmit the signs.

18. The method according to claim 15, wherein applying motion compensation includes applying bi-directional motion compensation implemented via multitexturing, the multitexturing including receiving color values from plural texture units, scaling the received color values, and applying the scaled color values to a fragment of the three-dimensional scenes to be rendered.

19. The method according to claim 1 wherein said operation of converting said three-dimensional scenes to be rendered into decoded video images comprises converting from a YUV color format to an RGB format.

12. The method according to claim 12, wherein performing said rasterization operation includes returning the set of pixels based on values of color coordinates of said compressed digital video images.

13. The method according to claim 12, wherein the returning includes returning the set of pixels based on the values of color coordinates even if the values of color coordinates are negative.

20. A system for decoding a stream of compressed digital video images, comprising:

a graphics accelerator including;

reading means for reading said stream of compressed digital video images;

creating means for creating, starting from said stream of compressed digital video images, three-dimensional scenes to be rendered; and

converting means for converting said three-dimensional scenes to be rendered into decoded video images.

21. The system according to claim 20 wherein said graphics accelerator is configured as a pipeline.

22. The system according to claim 20 wherein said graphics accelerator is selectively switchable between operation in a graphics context and operation for decoding said stream of video images.

23. The system according to claim 20 wherein said graphics accelerator is controllable in the operation for decoding said stream of compressed digital video images via a set of Application Programming Interfaces.

24. The system according to claim 23 wherein said set of Application Programming Interfaces comprises standard APIs for operation of said graphics accelerator in a graphics context.

25. The system according to said graphics accelerator claim 20 wherein said reading means comprises a stage for submitting said stream of compressed digital video images to at least one operation chosen in the group constituted by:

header decoding;

inverse variable-length coding;

zigzag scanning;

obtaining of quantized data regarding the blocks; and

decoding of the motion vectors.

26. The system according to claim 20 wherein said reading means comprises a stage for deriving from said stream of compressed digital video images the data on frame dimensions and frame number, type of images, and type of macroblocks.

27. The system according to claim 20 wherein said creating means comprises a geometrical stage for creating said three-dimensional scenes to be rendered as meshes of primitives.

28. The system according to claim 27 wherein said geometrical stage is configured for creating said three-dimensional scenes as meshes of two-dimensional geometrical figures, such as triangles.

29. The system according to claim 27 wherein said geometrical stage is configured for creating said three-dimensional scenes to be rendered as meshes of primitives contained in a Vertex Buffer Object.

30. The system according to claim 20 wherein said graphics accelerator comprises a geometrical stage for creating said three-dimensional scenes to be rendered via inverse quantization and inverse discrete cosine transform.

31. The system according to claim 20 wherein said creating means comprises a rasterizer stage for returning, starting from said meshes of primitives, a set of pixels representing prediction error frames of said decoded video images.

32. The system according to claim 31 wherein said rasterizer stage is supplied with values of color coordinates of said compressed digital video images.

33. The system according to claim 32 wherein said rasterizer stage is supplied with values of coordinates even if they are negative.

34. The system according to claim 20 wherein said converting means comprises a texture stage for converting said three-dimensional scenes to be rendered into decoded video images by combining said three-dimensional scenes to be rendered with motion data of said compressed digital video images and stored reference images and applying a function of motion compensation.

35. The system according to claim 34 wherein said texture stage is supplied with motion data without sign one a first channel and with the sign on a separate, second channel.

36. The system according to claim 35 wherein said texture stage is provided with a transparency data channel, which is used as said second channel for transmission of the sign.

37. The system according to claim 34 wherein said texture stage is configured for bi-directional motion compensation implemented via multitexturing, for scaling a value recovered from the image, adding the scaled value to a current fragment value.

38. A computer-readable medium having contents that cause a graphics accelerator to decode a stream of compressed digital video images according to a method comprising:

reading said stream of compressed digital video images using the graphics accelerator;

creating, via said graphics accelerator, starting from said stream of compressed digital video images, three-dimensional scenes to be rendered; and

converting said three-dimensional scenes to be rendered into decoded video images.

39. The computer-readable medium according to claim 38, wherein the method comprises controlling said graphics accelerator during operation for decoding said stream of compressed digital video images via a set of application programming interfaces.